[SPARK-28024] Incorrect numeric values when out of range

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.0.2, 2.1.3, 2.2.3, 2.3.4, 2.4.4, 3.0.0
• Component/s: SQL
• Labels: correctness
• Target Version/s: 3.0.0
• Assignee: Unassigned
• Reporter: Yuming Wang
• Created: 12/Jun/19 10:19
• Updated: 4 days ago 14:20

Description

Spark on master at commit de00ac8a05aedb3a150c8c10f76d1fe5496b1df3 with set spark.sql.ansi.enabled=true; as compared to the default behavior on PostgreSQL 16.

Case 1:

select tinyint(128) * tinyint(2); -- 0
select smallint(2147483647) * smallint(2); -- -2
select int(2147483647) * int(2); -- -2
SELECT smallint((-32768)) * smallint(-1); -- -32768

With ANSI mode enabled, this case is no longer an issue. All 4 of the above statements now yield CAST_OVERFLOW or ARITHMETIC_OVERFLOW errors.

Case 2:

spark-sql> select cast('10e-70' as float), cast('-10e-70' as float);
0.0	-0.0

postgres=# select cast('10e-70' as float), cast('-10e-70' as float);
 float8 | float8 
--------+--------
  1e-69 | -1e-69 

Case 3:

spark-sql> select cast('10e-400' as double), cast('-10e-400' as double);
0.0	-0.0

postgres=# select cast('10e-400' as double precision), cast('-10e-400' as double precision);
ERROR:  "10e-400" is out of range for type double precision
LINE 1: select cast('10e-400' as double precision), cast('-10e-400' ...
                    ^ 

Case 4:

spark-sql (default)> select exp(1.2345678901234E200);
Infinity

postgres=# select exp(1.2345678901234E200);
ERROR:  value overflows numeric format 

Issue Links

• is related to: SPARK-44444 Use ANSI SQL mode by default
• relates to: SPARK-23179 Support option to throw exception if overflow occurs during Decimal arithmetic
• relates to: SPARK-26218 Throw exception on overflow for integers